LOG taper faders turned into pseudo-LIN taper?????

[mikee]

Hi all.

In my local council recycling center, I found a 16 channel pro mixer desk in a bad state of repair.

Noting that it had lots of excellent Alps 100mm 10K Ohm Log faders, (plus lots of 100K Lin center-detented pots etc etc) I thought they might be good in a Mbox64.

Problem is that they are Log law tapers, and it seems that Mboxes don't like log pots. (What is the effect by the way?)

As I had at least 18 faders, I decided to use two faders arranged in parallel, with one fader inverted. The end result is a fader of 5K Ohm with what seems to be a pseudo-linear taper.

I know this is awfully wasteful, and I also end up with 5K pots rather than the standard MB 10K, but what the hell, and my power supply (being from a pc, minus noisy fan) can handle the extra bit of current easily.

Anyway, they work for me  ;)

I hope this input is helpful in showing what might be possible when faced with difficult situations.

Cheers all

Here's how they are wired up:

pots.JPG

pots.JPG

[lylehaze]

(What is the effect by the way?)

Our hearing sensitivity follows a log curve, so most pots that handle audio directly are (an approximation) of log curve. If you were to use a linear pot as an audio volume control, you would find most of the range of rotation wasted, and a small range where most of the useful settings are.

On the other hand, when we make "control circuits", like the MB64, we usually want the data to be a fair description of the position of the knob, so we use linear pots, and the voltage is a straight description of the knob position. If you used a log pot with an MB64, you'll find that most of the number range comes from one end of the pot, with MUCH slower response at the other end.

When I made the MBMixer, I scaled the knob data directly into dB data for the PGA chips. The result was the same as if I had used linear pots for audio material. It was necessary to add functions to scale from Linear to Log (and back again). The source in PIC Assembler is posted with the project, but it's NOT a great way to adapt the "wrong" types of pots, as resolution is lost in each conversion.

If you're the visual type, you can see Excel charts that I used to create the log conversion tables. They are in the MBMIxer wiki, or

http://www.midibox.org/dokuwiki/lib/exe/fetch.php?media=pga:logcurve.zip

Have Fun,

LyleHaze

[mikee]

Thanks for the quick response LyleHaze, I understand what you say.

The method I used for my spare log faders, although wasteful, actually straightens out the log curve to a great extent, thus allowing the new faders to function in a less cramped manner, than the original log-law faders.

The problem was that I had a bunch of 100mm faders, rotary pots and switches stripped from an old Allen & Heath SR416 pro mixer, and as  these components are pro-quality Alps components, I felt it was worth while attempting to use them, albeit after some additional practical modifications.

I also have Alps a bunch 100K lin center-indented rotary pots, (each fitted with a colour-coded knob) and I use each to drive an emitter follower, such that the resultant output voltage ranges from 0.3V to about 4.8V, this range, I am sure will be adequate to control various synth pots.

There are also some Alps non-indented 100K log rotary pots, and in this case I added a 12K resistor between +5V and the wiper, the result is a marked improvement over the Logarithmic curve, and again a component has been utilized, rather  than thrown away.

I must stress however, that I have yest to get a core module up and running, so I don't know what the actual effects of my mods will be, as I have only used a meter to test the things out. Your experience of the MBox and the effects of using log pots is helpful to me.

As for the switches, they are Alps dpdt latched push switches, of quality gold beryllium contacts, fitted with nice little caps. Why waste them? So I use one pole to feed a led, and the other pole feed a little circuit, which provides a negative-going pulse each time the switch is pressed, this is fed to the DINX4 inputs, and hopefully will be interpreted as a 'normal' button pressed to provide a logic 0.

When the switch is depressed (ON) the led lights up and a logic 0 pulse is provided. When the switch is depressed again (OFF) the led turns off, and another logic 0 pulse is sent. This at least will save on a load of chips and DOUTX4 boards, and if the serial read/write routines in the core are disabled, might help reduce the load on the chip, and perhaps leave the extra space for other stuff.

(An added benefit of using such latched switches is that they are able to drive higher currents and voltages than the chips on the DOUTX4 boards, which opens the door to using more interesting indicator lamps).

By the way, I understand the problems which might arise when using 100K pots rather than lower resistance 10K ones, but I wonder if anyone has actually experienced the effects of using such high values in practice. Would some decent screening help the matter? I only ask because there must be plenty of people with high-value pots, but unable, or unwilling to utilize them.

Thanks again LyleHaze for your interest, Ive downloaded the graphs and will look at them later.

Regards all.

[lylehaze]

It sounds like you are well along.

I'll add a  few comments, but please don't take anything I say as discouraging..

I'm just point out some details you are probably already aware of.

Sometimes a second train of thought is useful, or sometimes it just makes

a great train wreck. :-)

I use one pole to feed a led, and the other pole feed a little circuit, which provides a negative-going pulse each time the switch is pressed, this is fed to the DINX4 inputs, and hopefully will be interpreted as a 'normal' button pressed to provide a logic 0.

For the DINx4 side, You only need to short the input to ground. There is a pullup resistor in the DIN module. debouncing is done in software. Also, if you actually create a "negative-going" pulse the DIN may be damaged, as input voltage levels must stay between 0 and +5 volts. (you probably didn't mean that it would go below ground, but I'm covering all bases here!)

Logically, LEDs that display button action are cool, but often with MIDI controls there are multiple control origins (your panel, a window on the PC, a stompbox on stage, etc..) If MIDI ties them all together, sometimes it's nice to see if some OTHER source is pressing the same button.. This would be an argument for controlled LEDs..

By the way, I understand the problems which might arise when using 100K pots rather than lower resistance 10K ones, but I wonder if anyone has actually experienced the effects of using such high values in practice. Would some decent screening help the matter? I only ask because there must be plenty of people with high-value pots, but unable, or unwilling to utilize them.

I have not played with that, perhaps someone else has. Sounds like a great experiment..

Being a hardware type, if I wanted to use those pots, I'd just add op-amp buffers if there was a problem, but that may not be necessary. I seem to fix everything with Op-Amps.. But that's just me. :-)

Have Fun,

LyleHaze

[SLP]

As for the switches, they are Alps dpdt latched push switches, of quality gold beryllium contacts, fitted with nice little caps. Why waste them?

Because if you mix the beryllium with some americium (you can scratch that from a smoke-detector), you can make your own neutron source.

which is pretty awesome IMHO.

[lylehaze]

I'll have to wait until tonight to be sure, but I think that last post glowed in the dark, just a little bit. :-)

[nILS]

Hooray for the atomic boyscout!

[/tilted/]

I don't think I need any more neutrons.

At last count, I had the correct number.

...whatever that is.

[lylehaze]

"Last week, I lost an Electron" I said to the bartender...

"Are you sure?" he asked, seeing the despair in my face..

"Sure?", I said.. " I'm POSITIVE!"

[mikee]

Yeah, thanks for the input all.

Thanks LyleHaze. I did mean pulling down from +5V to ground (0V), which must, of course be a negative-going direction, its just old Logic speak I'm afraid. I was brought up to believe that moving from zero to a higher potential is positive-going, and from pos to a lower potential, neg going. Of course moving from zero in a negative-going direction will see the end of any standard chip, if applied to an input. It's all relative.

Your thoughts on the leds etc made sense.

The idea of using op amp followers (resistance converter-buffers) came up a while back, as I have lots of dual jfet input op amps on the same scrapped mixer, and I am still toying with the idea. Getting the supply voltages correct - so as to have a control output of 0V to 5V - is a bit of a pain, especially if trying to eliminate the use of dual polarity supplies.

Its all open to happy experiments though, and that's what its all really about.

Liked your graphs. Never used those packages before, I tend to employ Visual Basic to calc and display such data.

Regard all.

(My name's Bond, Co-valent Bond, as Enrico Fermi might have said.)

[SLP]

Hooray for the atomic boyscout!

nah, he doesn't look too healthy actually...

[img width=150]http://nicedeb.files.wordpress.com/2008/07/boyscout.jpg

[nILS]

Hmm, he looks just peachy to me.